[unrev-II] Patents and licenses

From: Lee Iverson (leei@ai.sri.com)
Date: Fri Jul 21 2000 - 13:00:01 PDT

  • Next message: Eric Armstrong: "Re: [unrev-II] Patents and licenses"

    I think we can resolve this all pretty quickly, I hope.

    I see three potential candidates for our license:

    1) The BSD/Apache license.
    2) The Mozilla Public License 1.1
    3) The GNU Public License 2.0

    These are actually in order of least to most restrictive. The BSD
    license basically says little more than "we retain copyright and do
    whatever you want with it". It does not actually contain any
    significant restrictions on modification or redistribution of the
    covered code.

    Both the MPL and GPL require source code to continue to be made
    available and that source code of modifications made to be available
    to end users. The major difference is that the GPL does not allow
    inclusion of the code in other systems which are not entirely released
    under an open source license. In other words, the GPL requires that
    any system that derives from GPLed software must allow an end-user
    free access to its entire source code (i.e. it must essentially be
    GPLed itself). The MPL merely requires that the MPL-licensed package
    itself and not necessarily other derived code be redistributed with
    source code including any modifications that the distributor has made.

    Now, as regarding patents. The GPL effectively tries to limit the
    consequences of patent claims against GPLed software by disallowing
    redistribution of software found to have been in conflict with the
    licensing obligations of a software patent. The MPL goes somewhat
    further and actually asserts that both the original owner of the code
    and all contributors explicitly allow royalty-free access to patented
    inventions *within the context of using the covered software* simply
    by releasing it under the MPL.

    Now, for my analysis:

    1) No licensing scheme will prevent someone from trying to patent
       ideas that we are trying to develop. The only real defense against
       that is public disclosure of *all* of the ideas, implementation
       strategies, aggregations of functionality etc. that we are
       pursuing. In patent terms, we want to produce a complete record
       that would be usable within a prior art database.

    2) The most important way in which the OHS can win is via ubiquity.
       What we really want is for our technologies to become as widespread
       as possible. I would argue that this cannot happen as long as it
       is not possible to use our work in commercial software, and by this
       I mean Microsoft Word, AppleWorks, Rational Rose etc. This, I
       would say, leaves the GPL out since it basically dictates a
       business model for software distribution.

    3) The concern over hijacking is *not* addressed by the BSD/Apache
       licenses, and Brian Behlendorf has essentially argued that the
       protection is more of a community and practicality issue than
       anything. While I buy that up to a point (it is kind of stupid to
       adopt an open source component and then proprietarize it and thus
       lose many of the benefits its being open source), I believe that
       argument doesn't reflect the potential for abuse by companies which
       can out-scale the open source community by virtue of their sheer
       size and market penetration (here I'm talking Microsoft and Oracle
       here as examples).

    All that said, I particularly like the combination of the MPL and GPL
    strategy. By licensing under both and allowing the potential adopter
    of you technology the option of choosing which one to live by, you can
    have the strengths of both without the downside of either. Commercial
    developers have the opportunity to adopt without the free license to
    corrupt (as per the MPL) while truly altruistic developers can
    aggregate with their own GPLed components and hope for the viruslike
    consequences down the line.

    Moreover, the possibilty for this aggregation is explicitly recognized
    by the Mozilla family (see http://www.mozilla.org/MPL/boilerplate-1.1/)
    and probably makes it easier for us to inject ourselves into that
    community as well.


    Lee Iverson SRI International
    leei@ai.sri.com 333 Ravenswood Ave., Menlo Park CA 94025
    http://www.ai.sri.com/~leei/ (650) 859-3307

    Need technology solutions for your business?
    Respond.com will Help!

    Community email addresses:
      Post message: unrev-II@onelist.com
      Subscribe: unrev-II-subscribe@onelist.com
      Unsubscribe: unrev-II-unsubscribe@onelist.com
      List owner: unrev-II-owner@onelist.com

    Shortcut URL to this page:

    This archive was generated by hypermail 2b29 : Fri Jul 21 2000 - 13:09:21 PDT